The fundamentals

of flame treatmentprocess
and its main parameters

November 2005 STEFANO MANCINELLI

esseCI SRL
Abstract – In this article flame treatment technology is focused from a process point of view. After a
general review on flame chemistry and flame profile and stability, are analysed main flame treatment

parameters, their meaning and their interdependence. Also several reference curves are provided to

the reader more involved in flame treatment, or who desires to deepen this matter, to orient better

himself in finding best flame treatment parameters values. For how this paper is conceived, it is so

particularly useful to BOPP line process engineers and technicians. An other article will follow,

describing unique features of a polypropylene flame treated surface and its comparison with other

surface treatment technologies, and presenting the wide range of flame treatment web applications.

1. INTRODUCTION

Flame treatment is an industrial process used to improve wetting and adhesion properties of
polyolefin films (BOPP - Bioriented Polypropylene; OPP – Oriented Polypropylene; PE, PET,
PS, etc.) and components, such as automobile body parts(bumpers, dashboards, headlights,
etc.) and blow-molded bottles.
Polyolefin materials, and in particular PP (polypropylene) have many good properties as: low
cost; can be worked and shaped quite easily; can get good mechanical properties, if properly
worked; are very good electric insulators.
Anyway they are apolar on their surfaces, which are characterized with very poor energies.
This is the reason why they need to be treated, in order to make possible their coating with
inks, paints, adhesives, metal, and other materials typically coupled with polyolefins, in
industrial applications as flexible packaging or automotive production.
Activation of polyolefin surfaces by flame is realized by means of two actions:
1. breaking of Carbon – Hydrogen links along the polymer surface, thanks to flame high
temperature, developed by the combustion process.
Theorical temperatures, reached using C 1–C3 hydrocarbon combustibles, are between
1700 °C and 1900 °C. As first step in the surface oxidation process, hydrogen
abstraction is far the more likely one, in comparison with breaking of a C – C link along
the macromolecular backbone.
Links C–C type are infact shielded, from radicals external attacks, by means of
hydrogen and metyl groups surrounding the molecule backbone. In addition, the
mobility of radicals –C type, coming from an eventual scission of C–C links, is really
reduced (because of radicals –C dimensions), so high is the probability of a
recombination, after the scission, between radicals –C and C-.
2. Insertion of oxygen based groups – contained inside the flame area - in
correspondence of broken links points, along the macromolecular chains.
The oxygen so transferred to the polymer surface acts as a bridge between the
polymer itself and the second material to be coupled with it.
2. FLAME CHEMISTRY
In order for flames to propagate, their reaction kinetics must be fast; i.e. the mixture must
be explosive.
Premixed laminar flame - in which the fuel and the oxidizer are thoroughly mixed prior to
combustion - is produced by radical/chain reactions occurring in a combustion system,
formed by an oxidizer (generally air) and a combustible (in a solid, liquid or gaseous state).
Here will be considered just the last case, being only gaseous combustibles typically
hydrocarbons (natural gas, methane, propane, LPG, etc.) – used for polyolefins surface flame
treatment .
Main steps of combustion radical reactions are the following:

(1) M → R ° ; chain initiation step

k1

(2) R ° + M →
k2
αR ° °+ M ' ; chain branching step
(3) R ° + M →
k3
R ° + P ; propagating step forming product
(4) R ° + M →
k4
P ; propagating step forming product
(5) ° " ; termination step forming product
R → P

Where: M, M’ = reactant molecules

R° = different radicals species = OH, O2, O, HO2, H (RADICAL POOL)
P, P’, P” = reaction products
α = MULTIPLICATION FACTOR
k1, k2, k3, k4, k5 = reaction rates

The chain initiation step (1) is endothermic and quite slow, it is not important in determining
the explosive condition, nor is it important in determining the products formed.
Its duty is to provide the radical that develops the radical pool of reaction (2). Radicals are
chemicals species characterized by high reactivity and mobility.
The combustion system undergoes, then, to exothermic reactions and, if sufficient heat is
not removed from the system itself, it becomes self – heating.
Since the rate of reaction, and thus rate of heat release, will both increase exponentially
with temperature (Arrhenius mode), the reactions rapidly run away; that is, the system
explodes – step (3) of the reaction: the propagating step.
k4 + k5
When α > α CRIT = 1 + , the explosion condition is reached by the combustion system
k2
and , if the mixture is inside its flammability limits (i.e. it has a right composition) and
within its explosive conditions (i.e. within right pressure-temperature boundaries for a
certain composition) the flame is generated and can propagate.
At this point flame is a subsonic combustion wave (running at a speed around 40/45 cm/s in
case of systems air/hydrocarbons). According to the formula representing α CRIT, the higher
is the rate of chain branching step (k2) and the lower the rates of termination steps (k 4, k5),
the more likely will be the explosion of the combustion system.
Termination steps are determined by radicals recombination (4) or their contact with cold
surfaces of reaction vessel (wall destruction) (5).
α has a value between 1 and 2, since during these steps, from one radical, depending on the
chain branching reaction (see below), one or two radicals can be formed.
Chain branching step produces a radical pool, according to the following reactions:

(6) H° + O 2 → O ° + OH °
(7) O ° °+ H 2 → H° + OH °
(8) H 2 + OH ° → H 2O + H °
(9) O ° °+ H 2 O → OH° + OH°

The sequence [Eqs. (6) – (9)], that is typical of H 2–O2 combustion system, but can be found
in every hydrocarbon - O2 combustion system, is of great importance in the oxidation
reaction mechanics of any hydrocarbon, in that it provides the essential chain branching and
propagating steps as well as the radical pool for fast reaction to occur, so for flame to
propagate.
It is this radical pool that develops the oxygen based groups inside the flame, used to
activate the polyolefin surface (2nd factor of action of the flame).

2. LAMINAR FLAME PROFILE

A premixed, laminar flame (in which fuel and oxidizer are thoroughly mixed prior to
combustion) as the one used in esseCI srl systems, does not have same temperature in
every its point, nor the same radicals species concentration.
Flame presents a temperature profile, according to the following drawing:
 FIG.1 (Characteristic laminar Bunsen
flame)
Three regions can be recognized:

1) Preheating/Precombustion zone, known as “dark zone” (blue colored in Fig.1):

This is the “coldest” region in the flame, where the premixed gases haven’t reached their
explosion condition. Its colour is dark, almost black.
This area, for what we’ll see below, is very rich in hydrogen radicals, that, recombining with
oxygen radicals, diminish their number, so inhibiting combustion process.
This zone is, for this reason, known also as reducing zone.

2) Luminous zone (red colored in Fig.1)

This is the hottest flame region, where combustion system highest temperature is reached
(in theory, around 1700 °C for natural gas-based flames and around 1900 °C/2000 °C for
propane–based flames).
In this zone reactants concentration, quite steady in the dark zone, drops significantly, and
combustion reaction and heat release take place. This zone is around 1/1,5 mm thick.
It’s also the region at highest oxygen radicals content.
The colour of this zone depends on fuel-air ratio: a deep violet radiation, with flame
becoming almost transparent if gas quantity is reduced, is produced when the mixture is gas
lean (due to excited CH radicals); a green radiation appears, instead, when the mixture is
gas rich (due to excited C2 molecules) and, at the bottom of the Jonoflame flame, in esseCI
mixture generators, appears a mist surrounding the flame itself, caused by excess gas
burning. When the mixture is very rich in gas a yellow radiation generally appears, produced
by carbon particles formed. The observation of colour of the flame, inside the Jonoflame
device, can be of great help to the flame treatment plant operator to keep right mixture
composition working conditions.

3) Post combustion zone (orange colored in Fig.1)

This zone is the largest of the three characterizing a laminar flame. Temperature is still
high here, since takes place, partially or completely, last reaction of combustion system, i.e.
the one converting CO in CO2 (that is highly exothermic).
This zone is intermediate, between the two before seen, in terms of temperature value and
oxygen radicals concentration. The high temperature burned gases generally present a
reddish colour, given by water vapour and carbon dioxide (combustion products with
hydrocarbon combustibles) radiation.

The existence of a profile in a laminar flame is mainly caused by convective flows

phenomenons (of unburned gases toward the flame zone) and by radicals diffusion, in the
opposite direction, from luminous zone into the dark zone.
In this zone hydrogen radicals, as already said, tend to combine with oxygen radicals,
forming HO2 radical and then, hydrogen peroxide (H2O2).
Because of low temperatures in the dark zone, the hydrogen peroxide doesn’t dissociate
and, by convective flows, is moved toward the reaction zone (luminous zone) of the flame,
where it forms OH radicals, thanks to higher temperature here present, according to the
reaction:

H2O2 + M ⇒ 2OH + M

where M is an intermediate species.

This explain high concentration of oxygen-based (OH) radicals in the luminous zone and the
very high temperature here reached, being the reaction of formation of OH radicals highly
exothermic (around 85 kcal/mol, including also reaction of H 2O2 formation starting from
hydrogen).

3. LAMINAR FLAME SPEED AND STABILITY

We’ve already seen that, in condition of propagation for the radicals reactions, representing
the combustion system, flame is a subsonic wave of combustion.

Flame velocity, or laminar flame speed, is defined as the velocity at which unburned gases
move through the combustion wave in the direction normal to the wave surface.
Different theories have been developed to describe and quantify flame speed; one of the
simplest is the Mallard – Le Chatelier theory, that is based on the concept that the
controlling mechanism in flame propagation is the heat diffusion back (from the combustion
zone) through layers of unburned gas, considering the flame given by a series of waves,
adjacent each other, formed by unburned gas at always higher temperature, up to reach the
combustion condition. According to this flame structure and referring to fig. No. 2 (see
below), the theory assumes that a flame can be divided in 2 zones, separated at the point
where the next layer ignites: the region of conduction (zone I) an the region of burning
(zone II).

FIG.2

The theory starts from the assessment that the heat, conducted from zone II in fig. 2, has
to be the same to that necessary to raise unburned gas to the ignition temperature
(boundary between zone I and zone II). Assuming linear the slope of the different adjacent
gas layers temperature, from the enthalpy balance is obtained the following formula:

mc p ( Ti − To ) = λ
( Tf − Ti ) A ( 3.1) λ = mixture thermal conductivity;
δ
m = mixture flow;
A = cross sectional area,
through which passes the mixture;
Being m=ρ Au (3.2); δ = flame luminous zone thickness;
cp = mixture specific heat ;
Ti = mixture ignition temperature
and SL = u (3.3) To = unburned mixture temperature;
Tf = flame temperature;
ρ = mixture density;
 u = mixture speed;
SL = flame speed.
in the hypothesis of a steady flame.
Equation (3.1) then becomes;

ρ SL c p ( Ti − To ) = λ ( Tf − Ti ) / δ (3.4)

from which we can get an expression for flame speed:

SL =
( λ ) (T f − Ti ) 1
(3.5)
ρc p ( Ti − To ) δ

from formula 3.5 it can be seen that final flame temperature and flame speed are directly
proportional dimensions; the higher is the flame speed, the higher will be flame
temperature, so that it’s possible to speak of flame temperature in terms of flame speed.
Since flame reaction zone ( δ ) is related to the radical reaction combustion rate (RR), by
means of the formula:

SL (3.6)
∂=
RR

the following relation is true:

SL ∼ ( αRR)1/ 2 (3.7)

λ
where α = ρ C = mixture thermal diffusivity (3.7’)
p

Equation (3.7) demonstrates the relationship between combustion reaction rate and flame
temperature, so the higher is the first and the higher will be the second.
Substituting equation (3.7) in equation (3.6):

SL SL α
δ = ≡ ≡ (3.8)
RR sL α
2 SL

Using, in equation (3.8), a value, for α, evaluated at a mean temperature of about 1300 K
and considering as flame speed (SL) a value around 35/40 cm/sec., realistic for
hydrocarbon flames, eq. (3.8) gives out for δ a value around 1.0/1,5 mm, as thickness of the
luminous zone.
Eq.s (3.5) and (3.7) show that flame temperature, as flame speed (S L) and reaction rate (RR)
– we have seen that talking of one of these quantities is the same as talking of the others –
depends directly on α value: the higher is α value and the higher will be the flame
temperature, so the higher will be also combustion system (then flame treatment) yield.
Getting an higher temperature flame means to increase the quality of the combustion
system.
esseCI srl research efforts have been and will continue to be focused on this field, the one
of “enriched flames”.
A way followed to increase thermal diffusivity ( α) is to reduce the mixture specific heat
(see eq. 3.7’); this can be obtained substituting nitrogen in the comburent with an other
lower cp inert, as Argon or Helium. It is reported from literature that methane flame
speeds, in stoichiometric conditions, in helium air is around three times (∼ 125 cm/sec.) the
one with Nitrogen, while the one in Argon air is around double ( ∼ 80 cm/sec.) the one in
Nitrogen. As an important consequence of the flame speed definition, is the concept of
flame stability from a fluidic point of view. Flame represents, in fact, a dynamic balance
between flame speed (so combustion reaction rate – RR), directed towards the burner grid
and the speed of the mixture coming out from the burner (see eq. 3.3). This balance is an
indifferent one; the flame puts itself in the point where flame speed equals the mixture
flow one. This balance is modified to an other balance condition (indifferent balance) if
mixture speed changes (this happens when mixture flow value is changed) or if flame speed
changes (this happens when mixture composition, i.e. the air/gas ratio; is changed).
At the extremes of the whole range of flame positions there are 2 instability points:

1. BLOWOFF: considering a flame anchored to the burner grid, if mixture flow speed is
increased, flame tends to come away from the grid.
At this point two competing phenomenons happen:

1. the cooling effect exerted by burner walls on the combustion system is

reduced, less heat is taken away from it and the higher is the flame temperature,
i.e. the flame speed (SL). Depending on the values of mixture flow speed, flame
tends to comeback to a stable equilibrium position;

2. when flame goes away from the grid, atmospheric air spreads inside the flame,
producing a diluent effect and the reduction of flame temperature and speed. If
this phenomenon prevails , SL is reduced at a point that produces flame blow off.

2) FLASHBACK: this phenomenon happens when flame speed (S L) is higher than mixture
flow speed, because, for example, of a strong reduction in mixture flow.
In the following fig. 3 is well shown the relationship between burning (flame) velocity
and mixture velocity inside the burner. dp is the penetration distance, i.e. the distance
from the burner walls, within whom, because of cooling and radical termination
effects exerted by the walls, SL can be assumed equal to zero.
The figure shows that if U1 is the mixture flow mean velocity through the burner
grid, there is no place where flame velocity is higher than U 1; the flame will
consequently blow out of the grid.
 U2 so represents minimum mean velocity before flashback occurs. U3 indicates a
region in which U3 < SL , so the flame does flashback.

FIG.3

An other point of view is possible in considering flame stability; it refers to the ability of
the air/gas mixture to sustain the flame, once the ignition source (spark) is removed.
The concept of flammability limits is then introduced; these are the limits in mixture
composition beneath whom flame doesn’t propagate.
They don’t have to be taken for explosive limits, that are pressure-temperature boundaries,
dividing conditions of slow and fast combustion reaction to occur.
Flammability limits are shown in literature both in terms of air and oxygen and are defined
by the competition between the heat of combustion reaction (depending on the reaction
rate) and the rate of heat loss by the flame, mainly because of radiation effects.
One limit is defined in the gas lean side of the mixture, and the other in the gas rich side.
Here below – table 1 – are summarized flammability limits for several very common
combustible species (in fuel volume percent):
TAB. 1

Lean Rich
Air O2 Air O2
H2 4 4 75 94
CO 12 16 74 94
NH3 15 15 28 79
CH4 5 5 15 61
C3H8 2 2 10 55

The table shows how lean limit doesn’t change considering as comburent, air or just
oxygen, since the excess oxygen, in mixture lean conditions, has the same
thermophysical properties as Nitrogen . Things go in a different way for the rich
 limit, where in oxygen, rich limit is significantly higher, due to flame higher
temperature reached in oxygen, thanks to the absence of any nitrogen.

4 - FLAME TREATMENT PROCESS VARIABLES

A typical flame treatment plant installed in a BOPP line – basically – is composed of the
following units, as the below fig. 4 shows:

1. Burner (water cooled), that produces a suitable flame for web surface treating.
2. Treater roll (water cooled), that supports the film, allowing its rewinding, while
cooling it, so preventing damages on film surface looking to the flame, due to
overheating.
3. Nip roll (suitably rubbered), having the duty to eliminate air layers/blisters
between film surface and treater roll, that compromise right thermal exchange
between the web and the treater roll itself.

FIG.4

Let examine main variables/parameters involved in a web flame treatment running:

A) Mixture flow

Mixture flow represents the quantity of energy supplied to the web, in order to get a
certain treatment on its surface, measured according to the specification ASTM D2578.
Generally speaking, the higher is line speed and the higher should be mixture flow.
Anyway, trend of treatment level (y – axis) vs. mixture flow (x – axis) for a certain line
speed is asymptotic toward an axis parallel to x – axis, as fig. 5 shows:
TREATMENT (dyne/cm)

Q1 Q2 MIXTURE FLOW (m3/h)

FIG.5

This trend explains the importance of grid width in the burner. Grid width infact
determines “a dwell time” of web under the flame (that is in the order of hundredth of
seconds) and so the power that can be transferred from flame to the web surface.
For a certain burner grid width there is so a threshold value in mixture flow (Q 1 in fig. 5),
working beneath whom (for example Q 2 in fig. 5) means to waste energy and thermally
stress web surface, since treatment level cannot increase further.
Basing on its 30 years experience, esseCI srl has developed several curves, relating, for
different line speeds, treatment level desired values to mixture flows used.
The curves are showed in below fig. 6 and fig. 7 and are so divided in two groups, one for
lower speeds (fig. 6) and the other for higher speeds (fig. 7).
These curves are to be considered as a general reference, valid for BOPP plane film, so as a
starting point for parameter mixture flow optimization.
In the two figures, mixture flow is expressed per burner meter, so to get total mixture
flow value, it’s necessary to multiply mixture flow value, got from the diagram
(corresponding, for a certain line speed, to the desired treatment level), by the flame
length (equal to burner length minus length of eventual flame trimmers used).
 TREATMENT LEVEL vs. MIXTURE FLOW CURVES
(AT DIFFERENT LINE SPEEDS)

54

52
SURFACE TENSION [dynes/cm]
according to ASTM D2578

50

48 v 150m/min
v 200m/min
46
v 250m/min
44 v 300m/min

42

40

38
10 20 30 40 50 60 70 80 90 100
MIXTURE FLOW [Nm c/h per burner m eter]

FIG.6

TREATMENT LEVEL vs. MIXTURE FLOW CURVES

(AT DIFFERENT LINE SPEEDS)

54

52
SURFACE TENSION [dynes/cm]
according to ASTM D2578

50

48 v 350m/min
v 400m/min
46
v 450m/min
44 v 500m/min

42

40

38
40 60 80 100 120 140
MIXTURE FLOW [Nm c/h per burner m eter]

FIG.7

B) Mixture composition – Air/Gas ratio

If mixture flow represents the quantity of energy supplied by flame to the web, mixture
composition is one of the two flame treatment variables involved with quality of energy
supplied to the material.
Maximum yield of a combustion system, so maximum flame temperature, is obtained in
correspondence of the stoichiometric point in mixture composition.
Stoichiometric composition is defined as the one in which there is exactly the right volume
of air to oxidize all the present gas, and the right volume of gas to be oxidized by the air
present in the combustion system.
So, in stoichiometric condition, there is no excess air nor excess gas; excess air and gas do
not participate chemically to the combustion reaction, then you will find this excess, within
reaction products, in the same amounts as it was in the reactants.
Anyway excess air or gas participates to the combustion system from a physical point of
view, since, depending on their specific heat value, these excesses take away heat from
combustion system, increasing their temperature, but reducing flame temperature.

FIG.8

Fig. 8 shows that maximum flame temperature is obtained in stoichiometric condition.

In fig. 8 is in fact represented flame temperature trend vs. mixture composition. This last
is expressed in terms of the parameter equivalence ratio φ , defined as the ratio of fuel on
air used, divided by the ratio of fuel on air stoichiometric. A parameter more frequently
used, in the automotive industry, to express mixture composition is lambda factor ( λ). λ is
defined as the equivalence ratio reciprocal (λ = φ -1).

So,

λ = (A/F) / (A/F)STOICH
and λ = 1 ⇒ mixture stoichiometric
 λ < 1 ⇒ mixture fuel– rich
λ > 1 ⇒ mixture fuel – lean .

In the introduction we have seen that web surface activation by flame is based both on
flame temperature and oxygen radicals – present inside the flame - actions.
Higher flame temperature means higher flame treatment yield. Anyway, mixture
composition generally used in esseCI srl plants is not the stoichiometric one, but a little
shifted toward fuel – lean side of diagram, in order to work with a little oxidizing flame, so
getting best compromise between high flame temperature and high oxygen radicals content.
Working with right mixture composition is, then, a key point for a flame treater correctly
running. This is the reason why esseCI srl mixture generators are equipped with a unit that,
in real time, monitors and controls mixture composition. This unit name is Jonoflame and its
P & I is shown in below fig. 9.

FIG.9

A sample (350/400 lph) of the mixture going to the burner is taken from the mixture piping
and used to feed a special burner, inside the Jonoflame, where it is burned. A
thermocouple, K type, placed at a fixed distance from the Jonoflame burner, measures
temperature of flame produced inside the Jonoflame itself.
For how Jono system is designed, with air false sucked inside the Jonoflame burner, higher
temperature read by the T/C corresponds to a gas richer mixture used.
So, Jonoflame system is a calorimetric one, in the sense that mixture composition is
expressed in terms of temperature.
Speaking of Jonoflame temperature means speaking of mixture composition.
T/C signals is sent to a controller driving gas control valve (where the regulation loop is
closed); the controller opens or closes the valve in order to zero the difference between
the Jonoflame temperature actual value and the set point one.
Other systems, based on the physic of Zirconium oxide and on massic flows measurements
have been developed by esseCI srl to back up the Jonoflame.
These devices form the JCS (“Jonoflame Control System”) line of products commercialised
by esseCI srl.
The importance of running with correct mixture composition has been underlined several
times during esseCI srl assistances to the customers and is showed on fig. 10, where it’s
possible to see that, leaving all the flame parameters (mixture flow, air gap, line speed) at a
constant value, and just changing the Jonoflame temperature value; i.e. mixture
composition, it’s possible to find a difference in treatment level, measured according to the
specification ASTM D2578, up to 10/12 dyne/cm.

TREATMENT LEVEL vs. JONO TEMPERATURE CURVE

(Qmix = 230m3/h; gap 4.0mm; v 160m/min; H2O roll = 27°C)

55
Treatment Level

50
[dynes/cm]

45
40
35
30
810 820 830 840
Jono temperature [°C]
FIG.10

Room conditions influence

The mixture composition parameters is heavily affected by room conditions, in particular by

values assumed by room temperature (T R) and by room relative humidity (HR%).
Variations in TR value and HR% one, produce a modification in mixture air/gas ratio in the
sense that, an increase in temperature and moisture values shifts mixture composition
toward fuel – rich side of composition, even if the Jonoflame temperature set – point and
the gas valve position didn’t move. This concept is well represented in fig. 11, where is shown
how mixture lambda value decreases when room temperature and/or relative moisture
increases. The diagram is referred to natural gas with relative density 0,59 and is based on
the hypothesis that the mixture is in stoichiometric condition at TR = 20 °C and HR% = 0.

Lam bda value vs. Room Tem perature at different relative hum idity values. Com bustion
in natural gas ( d=0.59). Stoichiometric at 20°C, RH% = 0

1,1

0,9
RH 0%
Lambda value

0,8 RH 25%
RH 50%
0,7 RH 75%
RH 100%
0,6

0,5

0,4
0 20 40 60 80 100
Room Temperature [°C]

FIG.11

So finding best mixture composition for a correct flame treatment running is not enough;
it’s necessary, in periods of high temperature and relative moisture, to balance the mixture
enrichment in gas, produced by room conditions, diminishing Jonoflame temperature set –
point.
esseCI srl has developed a device that can measure variations in mixture λ values in
correspondence of room conditions modifications.

C) Air Gap

With the term air gap is indicated the distance, in mm, between burner grid and treater roll
surface.
Along with mixture composition, air gap is the second flame treatment variable involved with
quality of energy supplied by flame to the web surface.
Since, as was indicated in chapter 2, flame has a profile, characterized by three distinct
regions at different temperatures and different oxygen radicals concentrations, it becomes
very important, in order to optimize the energy coming from the combustion of mixture
flow used, to put the burner (i.e. the flame) at the right distance from the treater roll.
Flame has to work in its luminous zone, the one at highest yield.
In the following fig. 12 is reported the air gap optimization curve, that, following the
attached instructions below the curve itself, allows to find air gap best operating windows
in correspondence of a certain mixture flow value.
This windows width is around 1.5 mm, since it corresponds to the flame luminous zone width,
where highest is flame temperature and radicals content and where the web surface has to
pass through. If flame works in its dark region, here flame temperature drops, hydrogen
radicals content is extremely high and they recombine with oxygen radicals, so the number
of these drops too and the treatment level obtained is low.
Working with air gap higher than the one corresponding to the luminous zone means, again,
not to work in optimum conditions, but the yield is, in this case, higher than in the case of
working in dark zone of the flame, since the post combustion zone is higher in temperature
and richer in oxygen radicals than the dark zone.
This is the reason why, in fig. 12 diagram, once best gap value has been found, the 1,5 mm
window is not centered on this value but is asymmetric, being larger on the higher gap side
than on the lower gap one (1.0 mm vs. 0.5 mm) – see below Note1 .
This curve emphasizes the air gap parameter dependence by mixture flow; increasing
mixture flow increases also air gap value. The 3 regions in the flame, since it is a dynamic
balance, are not fixed in position; they move according to mixture flow and mixture
composition variations. So, when mixture flow is changed, is very important to check if the
flame continues to work in its luminous zone. The instructions attached to fig. 12 diagram
underline also the difference, in working gap, between different esseCI grid types,
depending on their different patterns and net areas.

AIR GAP OPTIM IZATION CURVE

(METHANE 2.5 GAS; MIXTURE LAMBDA 1.028; GRID TYPE 0)

4,5
AIR GAP [mm]

3,5

2,5
1000 1200 1400 1600 1800 2000 2200 2400 2600 2800 3000
SPECIFIC SPEED [m/h]
FIG.12

Once working mixture flow is known (Q), to find the corresponding best gap is necessary to
use the formula below:
SPECIFIC SPEED [m/h] = Q / (useful flame length × grid width)
Where : - useful flame length [m] is the one obtained not considering the trimmers;
- grid width [m] is the width of the grid installed in the burner used (from12 mm
up to 34 mm).
 Known the specific speed is immediate to find, from the graph above, the optimum air
gap corresponding to the mixture flow used (Q).

Note1: Gap allowances = +1.0mm (in the sense of gap increasing) ; - 0.5mm (in the
sense of gap decreasing).
Note2: In case of grid no. 7, the graph above is still valid but the values on the y-axis
have to be reduced of 0.5mm.
Note3: In case you're using LPG gas instead of Methane, the graph above is still valid
but the values on the y-axis have to be reduced of 0.5mm.

Air Gap value is influenced also by mixture composition, even if in a lesser extent
than by mixture flow. This concept is well illustrated in fig. 13 below, where is
represented flame temperature trend for three different compositions,
(stoichiometric, poor, rich), for a certain constant mixture flow value.
The curves were obtained by several tests. Their parameters and devices are below
summarized:
1. Mixture flow used: 944 Nl/h.
2. Grid type/gross area: grid 0 / 5.8 cm2.
3. Air/gas flows measured by means of Bronkhorst massic flow meters model
F–201C-FGB 22 V and F-112AC-HDD 22V.
4. Flame temperature measured by means of a B – type thermocouple, able of
three degrees of freedom (one in vertical sense, representing the gap
variation; the other two horizontal, in order to explore all the possible areas on
the grid pattern).
5. Gas used: pure methane (99.95% content).
As we’ve already seen, the three curves trend underlines that:
1. maximum flame temperature is achieved in stoichiometric conditions (λ = 1,000);
2. taking as reference one of the three curves, it can be seen that flame temperature
drops faster – from point of maximum – going toward lower gaps than going toward
higher ones;
3. the diagram shows also that air gap variable is affected, although in a lesser extent
than by mixture flow, by mixture composition variations.
The green curve – the gas richest one – has its maximum temperature point
corresponding to the smallest air gap value, compared with the other two curves.
A fuel – rich flame is shorter than a fuel – lean one.
Anyway, gap variation caused by significant mixture composition variations, well
beyond the ones generally recorded on a flame treatment plant, are of the order of
0,1/0,2 mm, so lower than 0,3 mm, value that can be assumed as threshold between
significant (from a process point of view) and insignificant air-gap variations.
So, if air gap value is already well centred, no gap adjustments are necessary in
correspondence of a mixture composition (i.e. Jonoflame temperature) variation.
 FLAME TEMPERATURE vs. AIR GAP for different m ixture com positions

1380
LAMBDA 1.000
1370
LAMBDA 1.050

1360 LAMBDA 0.950

1350

1340

1330

1320

1310

1300
2 3 4 5 6 7 8
AIR GAP [mm]

FIG.13

D) Treater roll temperature

Treater roll gives support to the web, during rewinding, so allowing its on-line
treatment. The roll is motorized to maintain same peripheric speed as the one of the
film . Energy produced by flame is in part evacuated through the hoods of an exhaust
system and through the water circulating inside the burner; the remaining is
transferred to the web. Since just around 10 -6 of the energy produced is acquired by
web surface to increase its wettability, the main part of energy that is transmitted,
by radiant and convective heat exchange, to the web, passes through it (producing a
sudden and time concentrated warming of film surface exposed to the flame itself)
by heat conduction and finally arrives to the treater roll, that is internally cooled.
On the average, around 50% of the energy produced by combustion is taken away by
water circulating both in the treater roll and in the burner.
For this reason it is mandatory to warrant, during flame treatment a perfect
adhesion of film to the treater roll , absolutely avoiding air spaces between them,
that dramatically reduce their heat exchange.
Air space absence is obtained working with the correct coupling between treater and
nip roll. In this condition, heat exchange coefficient between web and roll can be
assumed infinite and film surface – the one in contact with roll surface – presents a
temperature almost equal to the one of the roll surface, from the moment of contact
up to when it leaves this.
EsseCI srl treater roll – named SCR - is formed by two coaxial cylinders and a sheaf
of parallel piping between them, running along these cylinders generatrix, that allows
to reduce water passage section, so increasing water speed inside the roll and its
heat exchange capacity (see FIG.14). Water runs, in its outward journey, in several
parallel channels outside the piping and comes back running inside them.

FIG.14
The thermic power P, referred to each burner meter, conduction transmitted from
film to roll surface – equal to the one transmitted by forced convection, from roll
surface to the water at its insides - can be expressed as:

P = KA∆T ; [W/m]
where:
K = roll surface/water global heat exchange coefficient [W/(m2 °C)];
A = roll external surface [m2];
∆ T = medium temperature gradient between roll external surface and water
temperature (average between roll inlet and outlet values).

Assuming planar the surface through which heat is exchanged, the following formula
can be used for K:

1 S 1
= + + Rd ; [m2 °C/W]
K A × λm h
where:
Rd = piping pollution factor (takes in account piping encrustations and water hardness;
≅ 0.000175 m °C/W for treated tower water);
2

λm = conductivity of material forming roll external surface (50 W/(m °C) for steel);
h = water/roll surface thermal exchange coefficient [W/m2 °C] (≅ 3900 W/m2 °C for
water at 30°C)
S = external cylinder thickness
 from the above, K values for an esseCI treater roll have, on the average, values
between 1600 and 1800 W/m2 °C .
It is clear that treater roll temperature, to be chosen during flame treatment process, is
strongly influenced by, among the other variables, roll heat exchange coefficient.
Generally speaking, anyway, treater roll temperature has to be selected according to film
type to be treated:
1. in case of plain films best is to find, as treater roll temperature, the highest
possible, compatible with web surface treated good quality. So, a range of
temperatures between 26°C up to 45°C can be assumed as a good reference (with
lower values to be used with higher web thicknesses);
2. in case of thermosensitive films (as white opaque, cavitated, coex, cast, etc.),
attention has to be focused in avoiding any surface damage coming out from flame
treatment. This is the reason why best is to run the treatment with a treater roll
temperature being in the range 16 ÷ 26°C (with lower values to be used with higher
thickness). In this way is possible to get typical flame results (as superior surface
adhesion) without affecting other web properties (as sealability, flatness, optical
qualities, etc.), as we will see in next article.

STEFANO MANCINELLI

esseCI SRL, Via Flaminia Ternana n. 386 – 05035 Narni (Terni) – Italy

Phone: 0039.0744.726741; fax: 0039.0744.726953;

mailto: [email protected]; web: www.essecinet.com

REFERENCES

1. GLASSMAN I. – “Combustion”. Chapter 3 and 4. Academic Press Inc. –

N.Y. (1987).
2. BRADLEY J.N. – “Flame and combustion phenomena” Chapter 3. Methuen,
London (1969).
3. WESTBROOK C.K. and DRYER F.L. – “Prog. Energy Combust. Sc.” 10,1
(1984).
4. HOTTEL K.C.; WILLIAMS G.C.; NERHEIM N.M. and SCHNEIDER G.R. –
’’ Int. Symp. Combust.” Combustion Institute Pittsburgh – PA – (1965).
5. GIBBS G.J. and CALCOTE H.F. – “J. Chem. Eng. Data” (1959).
6. ZEBATAKIS K.S. – “U.S. Bur. Mines Bull. No.627” (1965).
7. CLINGMAN W.H. Jr.; BROKAW R.S. and PEASE R.D. – “Int. Symp.
Combust.” 24th . Williams and Wilkins, Baltimore (1953).
8. LEASON D.B. – “Int. Symp.Combust. 4th “ Williams and Wilkins,
Baltimore (1953).
9. LEWIS B. and Von ELBE G. – ’’Combustion, flames and explosion of gases“
2nd ed. Academic Press Inc. – N.Y. (1961).
10. LORI G. – “Aspetti chimico-fisici nell’attivazione di superfici mediante il
trattamento alla fiamma” (1985).
11. ANGELI A. – “Lo sviluppo e le prospettive del trattamento alla fiamma
polarizzata su vari tipi di film, carta ed alluminio” . 8th esseCI Meeting (1996).
12. LORI G. – “Il rullo di trattamento SCR”. 8th esseCI Meeting (1996).
13. STROBEL M.; BRANCH M.C; ULSH M.; KAPAUN R.S.; KIRK S.; LYONS
C.S. – “Flame surface modification of polypropylene film”. J. Adhesion
Sci.Technol. Vol.10, No.6 (1996).
14. STROBEL M.; SULLIVAN N.; BRANCH M.C; JONES V.; PARK J.; ULSH
M.; STROBEL J.M.; LYONS C.S. – “Gas phase modeling of impinging flames used
for the flame surface modification of polypropylene film”. J. Adhesion
Sci.Technol. Vol.15, No.1 (2001).